host range experiments) exist for assessing the potential direct effects of proposed biocontrol agents, to our knowledge, there are currently no methods available for quantifying and ranking potential non-target impacts of proposed biocontrol agents via indirect effects prior to their release, despite them being likely to be common. These indirect effects are more difficult to observe (or predict) because of the inherent difficulty in studying whole communities of interacting species and attributing causation to changes in potentially indirectly affected populations. the effect of one species on another mediated by a third species). direct effects), biocontrol agents have the potential to affect communities through indirect effects (i.e. However, in addition to directly affecting individual non-target species (i.e. Therefore, tools for assessing the risk of proposed biocontrol agents prior to their release are crucial for deciding whether a given agent should be introduced. Although biocontrol agents are often a more environmentally friendly means of suppressing pests than synthetic pesticides, they can have non-target effects on native species. Growing public concern about the harmful health and environmental effects of pesticides, combined with the rapid evolution of pest resistance to chemical control, mean that biocontrol agents are increasingly being advocated to suppress pests. Combining machine-learning and network approaches provides a starting point for reducing risk in biocontrol introductions, and could be applied more generally to predicting species interactions such as impacts of invasive species. Further, although our machine-learning informed methods could significantly predict indirect effects, the explanatory power of our machine-learning models for indirect interactions was reasonably low.
![list of biocontrol agents list of biocontrol agents](https://ars.els-cdn.com/content/image/3-s2.0-B9780128218433000143-f12-02-9780128218433.jpg)
This predictive ability depended on the generality of the interacting partners for KNN models, and depended on species’ abundances for both random-forest and KNN models, but did not depend on the source (habitat type) of data used to train the models. We found that random-forest models predicted host-parasitoid pairwise interactions (which could be used to predict attack of non-target host species) more successfully than KNN. apparent competition), and tested these predictions against empirically observed indirect effects between hosts. Finally, we used these predicted networks to generate predictions of indirect effects via shared natural enemies (i.e. Then, we tested whether the accuracy of machine-learning-predicted interactions depended on the generality or abundance of the interacting partners, or on the source (habitat type) of the training data. Here, we used two machine-learning techniques (random forest and k-nearest neighbour KNN) to test whether we could accurately predict empirically-observed quantitative host-parasitoid networks using trait and phylogenetic information. less abundant species, and across different habitat types is also untested for consumer-prey interactions.
![list of biocontrol agents list of biocontrol agents](https://www.anime-planet.com/images/characters/sajin-komamura-105.jpg)
direct effects) can be made equally well for generalists vs. Whether predictions of interactions (i.e. Independently, species traits and phylogenies have been shown to successfully predict species interactions and network structure (alleviating the need to collect quantitative interaction data), but whether these approaches can be combined to predict indirect impacts of natural enemies remains untested. The analysis of ecological networks offers a promising approach to understanding the community-wide impacts of biocontrol agents (via direct and indirect interactions).
![list of biocontrol agents list of biocontrol agents](https://www.mdpi.com/insects/insects-07-00070/article_deploy/html/images/insects-07-00070-g001.png)
‘biocontrol’) agents can have direct and indirect non-target impacts, and predicting these effects (especially indirect impacts) remains a central challenge in biocontrol risk assessment.